Noise is an inescapable feature of genetic networks, causing variation among genetically identical cells even under identical environmental conditions. Because expression is a multistep process involving many biochemical reactions, many recent studies have conjectured that natural selection can tune levels of noise to avoid fluctuations which could impair cellular function. One important parameter in tests for selection is the rate at which new variation in a trait is generated by mutation. This proposal aims to (1) estimate the rate at which noise and its effects on network activity evolve solely due to mutation, using the well-studied galactose network of Saccharomyces cerevisiae as a model system, (2) measure how robust the key positive feedback loop of this network is to regulatory mutations, and (3) use theory and simulation to analyze the evolution of a basic regulatory network under mutation and selection. Aims (1) and (2) will be pursued through in vivo monitoring of gene and network activity in cells exposed to varying levels of galactose carrying mutations randomly throughout the genome (1), or targeted to a particular network component (2). Mutant alleles play an important role in congenital diseases, and the accumulation of mutations is one of the features of cellular progression to cancer. In both of these cases, mutant alleles probably disrupt normal genetic network functioning, and so it is important to understand how genetic networks buffer the effects of mutations and how easily network function can be compromised. [unreadable] [unreadable] [unreadable]